meta.programmers.stackexchange.com/questions/4013/… has some stuff about it. Rachel's profile page has her summary. The site started out allowing a lot of questions it no longer does. It started very different for a little while, in fact, see meta.programmers.stackexchange.com/questions/3273/…. I'm not sure if the people leaving really want NPR again though, or some form of happy medium.

It's hard to answer without the danger of speaking for other people

Yeah I was reading the pierre meta

and then thorbjørn has also expressed his intention to leave

And DeveloperArt apparently has, and mostly Rachel. But I don't honestly know if the number of people leaving is higher than other sites. Or even the number of high reputation people.

@WorldEngineer But that's a false dichotomy. Programmers.SE shouldn't be like Reddit, but it doesn't have to be overly bureaucratic either. It can actually be fun too!

@WorldEngineer Ah. I see. Cool!

@WorldEngineer I just love Arqade's new skin. [But doesn't everybody?]

@Deco Correction - Everyone is not leaving P.SE. Pierre and DeveloperArt were big losses, and most of us (myself included) would like @Rachel to participate more often, but the community is still strong.

@WorldEngineer Right. I actually don't game at all anymore. EA Sports stopped making the Madden series for the PC, and their console programmers just haven't met my lofty standards (of course, I'm a programmer, so you'd think I'd have lofty standards ;)

@WorldEngineer Back in the day, I might have spent 10-12 hours per week playing the Madden and NHL Hockey series.

@WorldEngineer Oh?

@WorldEngineer Ha! Wow. I hear you.

@WorldEngineer HA! Yep. I might agree with that; but you seem like a decent guy anyhow.

@WorldEngineer: Of course these days, with all of the concussion studies and so forth, many people are rethinking the wisdom of letting their kids play football. [Even Tom Brady's dad!]

@WorldEngineer Ha! Yep! Wouldn't be too far from the American football concept.

@WorldEngineer Oh?

@JimG. No, there's definitely an exodus of sorts going on. Pierre and the other recent ones are just the current wave of them - but it's been going on for months. I've become less active here, too.

@Izkata Can you quantify the exodus? How many people have left?

@psr The "open source frameworks suck" question was closed again, your edit was great but the OP didn't feel like posting some examples.

Oh well, we tried ;)

@Rachel @psr @MadKeithV regarding reopening questions there is related feature request at MSO: ]

Add a “Review posts with reopen votes” review task

I put 50 bounty on it some time ago, it has expired since

I plan to put one or two more bounties on this request but

alone, I would unlikely be able to keep it float long enough to get implemented

you know successive bounties from same user go double each time as 50, 100, 200 rep

if it would be implemented I think it would be much easier to salvage posts that look worth reopening to you

I for one would be happy to actively use such a tool in review queue

it really looks unfair now that closing got so much easier with review close votes queue but reopening remains as cumbersome as it was

@gnat - I like that idea and have upvoted it and offered a bounty.

@psr - One question that I think is now good almost enough to be re-opened is programmers.stackexchange.com/questions/165364/… . Unfortunately I think the new user will never be back to look at that question again after the rapid closing because they probably have no idea that "closing" is a temporary state.

@gnat FYI there's a list of questions with re-open votes in the 10K tools.

@MadKeithV That's a problem, closed is such an ugly word, and nothing about it feels temporary. On top of it, users don't get any notification that one of their questions was closed (feature proposed and declined).

@YannisRizos - it's a shame really, it would be nice if there was a close (or reopen) option that would clearly mark the question as "could work with fixing" and notify the original poster. But that would probably still not help with those first-time users looking for a quick fix. On a side note: has anyone else noticed that the front page seems to have fewer closed questions than a week ago? ;-)

@MadKeithV Requiring registration seems to be working then?

@YannisRizos - it looks like it is. BTW, to make it explicit, I'm dogfooding here, trying to spread a positive vibe, and I see you're doing the same (cfr your edits to programmers.stackexchange.com/questions/165484/…).

@MadKeithV Re your answer there: While array values can be anything (and mixing values of diff types is perfectly fine), keys can only be int or string. And PHP will do some automatic type castings if it's anything else (see under Example #1), perhaps you'd want to point that out in your answer. The important gotcha is that although string keys are valid, if your string key is also a valid int, it will be autocasted to int without warning.

@YannisRizos in 10K tools, i see. this likely explains why Rachel did not notice that feature request at MSO. With her >10K she has an access to it and maybe doesn't know that possible "allies in reopening" under 10K simply don't have such a convenient access. It's exactly my case for example

@gnat Don't know if the feature will be build, till then searching for closed:1 is a good enough re-open queue... Do you really care if anyone else has casted a re-open vote?

@YannisRizos I typically check if there were edits accompanying reopen vote - basically reopen votes serve to me as a shortcut to review potentially salvaging edits

if there are no edits, I try tio figure myself what could make "re-opener" think that question could be improved - and if I can see it through, I don't hesitate to edit it myself :)

@YannisRizos - thanks for the heads up, and a self-smackdown for answering a question outside my knowledge area with a quick google, without understanding the full pitfalls.

Downvoted one of Jon Skeet's answers, and a second later a badge notification appeared. For a moment there I hoped for a super secret "done the unthinkable" badge ;) (turned out to be just a boring "nice answer" one)

@gnat Thanks, I went to that MSO question about the reopen queue and see I already have it voted up, so must have seen it at some point in the past. I do review the questions with reopen votes in the 10K tools fairly regularly (every week or two), however reopen votes expire so I know I miss some reopen attempts. And in many other cases I can't vote to reopen because I've already cast my reopen vote in the past, and can't recast it even if the question didn't get reopened

@Rachel it would be nice if you keep that question "on the radar" so to speak. I am not sure if my MSO rep will do up fast enough to allow keeping it hot with more than a series of two or maybe three bounties; would be great to have someone else involved. It has been quite an issue to me that it's so hard to catch and track reopen-candidate questions. Now that I work with close votes queue quite a lot, I really see how it could be done better for reopen

basically same design that works so well for closing, with minimal tweaking can be done for reopen queue

@gnat - Upvoted your bountied Add a "Review posts with reopen votes" question. It sounds like they plan to get around to it, for what that's worth.

@YannisRizos - So the "Skeet Shooting" badge is a myth?

Good lord... who would have thought creating a sports league season schedule algorithm would be so friggin hard...

WTF... graph theory?! i am in over my head

@mmyers hi :)

@voretaq7 hi

ok

I have a league with variable conferences, variable divisions of conferences and variable teams in divisions

The season is a variable length of games

What is this, the Play-Doh league? Are there no rules at all?

Half of the seasons games or the total number of teams in a division, whichever one is smaller, will consist of inner division games

Inner-division games should be random match ups

Any remaining games should be random teams in other conferences and divisions

If the total number of teams are even then everybody should have a full seaosn

If the total number of teams are odd then some team will end up with a random bye week

No team should have more than one bye week

@mmyers I know it sounds weird but our users all have wildly different league structure demands

we have to support just about anything they demand

my brain... it hurts... the pain...

Ah

Doesn't the NFL have a couple of full-time schedulers or something?

@mmyers :(

What's the range of league sizes?

@mmyers 0...n

everything...

is....

0...n

clearly 0 is a ridiculous number

This is still in design phase, and i have some pull on requirements, if certain aspects of this make it NP HARD then I can negotiate limitations.

this is what I have so far

I think if I can just sort through every team and make sure that each one has N games where N is season game length, then I will be okay...

uniqueGames can be uniquely identified by the two teams that are playing...

I'd probably start by special-casing leagues with fewer than 4 teams

@mmyers Premature optimization is the root of all evil

@maple_shaft Just to ignore the possibility for the moment, I mean. 4...n makes more sense than 0...n to me

@mmyers yeah I would work the problem with a reasonable floor and ceiling to start

@mmyers I suppose you are right... but realistically I don't see any of our clients only having four teams... it is one of those "maybe" requirements

are your random match-ups truly random? (I assume the answer is no -- true random would mean that a division w/ teams A B C D E could have half the games be AE and the other half be A[someone in another division])

@voretaq7 Hmm... well certainly not random for the list of all teams, but definitely random A-E sublist and not(A-E) sublist

So closer to "Half the games are comprised of distinct pairings without repetition until Gn > min(nTeams,nGames/2)" right?

@maple_shaft I don't think I understand your league rules, but last time I tried to do an algorithm for scheduling something I had some success by determining the most restrictive placements first, and working towards the least restrictive, with random being used anytime two elements were on the same level and competing for the same spot. Don't know if that makes sense or helps at all.... :)

(Obviously you have to do the "how many games will each team play?" bit first but that can be kicked as a separate problem)

@voretaq7 @Rachel that makes sense... it is actually how I started, but for some reason I got confused half way through... see my code above

mapping that onto a calendar though is a HUGE pain in the ass (but again separate problem - mainly handled by shuffling the order of games so you don't schedule team A to be in two different cities on the same day)

@maple_shaft sorry to spoil the fun, would you mind taking a look into a question with some slippery rollbacks? programmers.stackexchange.com/posts/165380/revisions OP tries to wipe out auto-inserted link to duplicate

I am going to be in the shower tomorrow morning and have a genius moment of clarity

@maple_shaft probably. That's how I solve problems like this :P

There should be a shower built in to every cubicle. Productivity would triple.

the graph theory approach is a sound one, but that's some deep magic compared to shuffling...

(and either you wind up with cyclic graphs or multiple trees with independent roots around each team)

@maple_shaft when in the shower, try not to forget that the law of gravity and periodic table are already invented. :) Keep focusing on scheduling games. If memory serves I've seen something about this problem in Skienna algorithms manual

@voretaq7 fxxx all that

... also I don't have to worry about locations and home and away restrictions :)

Every game is in the same place at the same time

((cyclic graphs naturally solve some of the same team in two cities problems, but you have to break / prune the cycles ; multiple roots don't need to have the cycle-finding code, but can put the same team in two matches on the same day))

@maple_shaft that's helpful, travel has to be one of the major headaches for non NFL schedulers

@maple_shaft Can a team play two games at once? :-)

and the Jets/Giants thing has to be hell for the NFL

(PLEASE say yes)

@voretaq7 no

@waxeagle ahem -- FUCK the NFL.

@voretaq7 :)

@maple_shaft OK, well then you have to do second-level shuffling to prevent that if you use my naive approach

@gnat wait what is happening? Did this guy just try to remove the Possible Duplicate link to unclose his question

EG:
Team A vs: C B D E
Team B vs: ? A ? ?

@gnat thanks for pointing that out... locked it... hopefully he gets the message

@maple_shaft I can't read really their mind, all I saw is that revision 3 they made was plain removal of auto-inserted link. edit summary is "deleted 249 characters in body", not particularly informative

I see - thanks

@gnat maybe he disagreed that it was a duplicate question or thought that by removing the link it would un-close his question?

lol

@voretaq7 @mmyers @waxeagle @Rachel Thanks guys! later!

@maple_shaft maybe. Or maybe not. My mind reading skills are not that mastered yet

@maple_shaft - Isn't your scheduling problem a valid question on Programmers? It's a real programming problem you face, you aren't looking for code, just an algorithm, and it's somewhat localized but not crazy localized.

I know it's the kind of question I would like to see more of, personally.

@maple_shaft - Have you thought the constraints through for funny cases? What if there are 4 teams in a division, 8 in the league, 200 game season. According to your constraints, 4 of the 200 games will be divisional, 196 with the other division.

@maple_shaft - Since you have few constraints, how about: At any given time everyone is playing in division or everyone is playing out. Keep track of how many times each pair of teams has played each other. When pairing, randomly pick a team, and randomly choose from among the eligible (i.e. (in/out) of division) opponents that team has played the least. Odd team out in division with odd number of teams goes into a temp division with all other odd teams out and follows same rule.